Identify the optimizer–load elasticity parameter

Identify the elasticity of compute demand to reductions in effective unit cost caused by data-center efficiency improvements, so that the net energy, carbon, water, and embodied-carbon effects of learned control can be measured rather than estimated from an unspecified parameter.

Background

CLEAR-DC introduces an elasticity parameter ε to represent induced demand: when efficiency lowers the effective unit cost of computation, demand may increase and offset direct facility savings. The paper’s net-benefit formulation depends on this parameter to distinguish direct savings from rebound effects.

The corpus does not provide data that identify ε. Consequently, the proposed framework specifies how the parameter should enter evaluation but does not resolve its value. Estimating it is necessary to convert the framework from an accounting specification into an empirical measurement of net benefit.

References

In particular, $\varepsilon$ is not identified by any dataset in this corpus; the definition makes its absence visible rather than supplying its value.

— Artificial Intelligence for Energy Optimization in Data Centers  (2609.03716 - Ullah et al., 3 Sep 2026) in Section III, subsection “Problem Formulation”; Section VII, subsection “Limits of the Evidence”